1. Field of the Invention
The present invention relates to a semiconductor memory device and a method of accessing it, particularly to a semiconductor memory device having a storage memory area and a working memory area.
2. Description of a Related Art
A conventional data processing system comprises at least a controller such as MPU, CPU, etc., for processing data and a memory device for storing data to be processed. Data has two types, one of which is necessary to maintain its information even after power is switched off, the other of which is used temporarily during data processing. The former data is stored in a non-volatile memory. The latter data is stored in a volatile memory to satisfy its high speed processing demand. The system engineer selectively use the two kinds of memories by storing the former in a storage memory and the latter in a working memory, based on the system to be designed. For example, a handy type information processing apparatus such as a mobile telephone, PDA, etc., use a flash memory as a storage memory and DRAM or SRAM as a working memory.
Regarding such a handy type information processing system, the miniaturization, weight saving, high-speed performance and diversity is demanded. To satisfy the demand, semiconductor memories such as DRAM, SRAM, etc. are developed.
However, there is some difficulty for system designer to use the two kinds of memory. For example, a flash memory requires as much as 1 μsec to store data, and thus, it takes a long time to write a control program into a mobile phone. This poses a problem, because, with the advent of high performance mobile phones in recent times, writing huge data such as a file of an animation film into a memory is required and actually tried even in those mobile phones, and this tendency is totally incompatible with said drawback of the flash memory, i.e., slowness in data writing. On the other hand, DRAM and SRAM have their own problem. DRAM does not have a bright prospect for high density integration because of its increasingly complicated procedures involved in the manufacture of its capacitor components. To ensure a required capacity, each capacitor component must have a certain height, and this hinders the progress of the device towards higher integration. To avoid this drawback, the capacitor component must be made of a material having a high dielectric constant, which will lead to an increased production cost. With respect to SRAM, each cell must include plural transistors, and thus to find a suitable method how to reduce a chip area becomes a problem. Thus, if one expects that DRAM or SRAM will serve as a memory through which he can handle a huge quantity of data at a high speed, he will meet a certain limitation sooner or later. Moreover, because the memory of a system consists of a non-volatile and volatile components, the production cost of the system is more increased than would be otherwise. Because access speeds to the two memory components are obviously different, management of the memory at large becomes difficult which complicates the designing of the system.